1. Who is the author?
Liu Xuefeng, male, doctor, associate professor and doctoral supervisor. 20 14 is an associate professor at Huazhong university of science and technology, and is currently an associate professor at the school of computer science, Beihang university. At present, the main research directions are:
1) artificial intelligence distributed algorithm: This paper focuses on how to realize the deep learning model (the application of federated learning in medical field) in which multiple medical institutions jointly train medical images.
2) Deep learning model with prior knowledge: Focus on how to integrate doctors' domain knowledge with data-driven deep learning model to solve the problem of less medical data.
3) Intelligent perception technology: combine the signal processing technology in the mobile field with deep learning to realize ubiquitous intelligence. Including positioning technology, human body state perception, etc.
4) Image recognition and natural language processing: The multimodal information (including images, sounds and natural language) fusion model plays a role in practical applications (including public opinion analysis, recommendation, emotion prediction, EEG signal processing and gesture recognition).
There are many proper nouns introduced above, which are difficult for laymen to understand, but we can get some exact information that the author has in-depth research on artificial intelligence computers. When it comes to computers, we must mention the network, which is bound to be linked with netizens. Now our life is surrounded by the Internet and various algorithms. We only talked about one thing with our colleagues at noon, and when we turned on our mobile phone in the afternoon, we were pushed a similar little video, which made me feel that I was being monitored by my smartphone. I heard that this is a computer algorithm.
Second, mathematics? What does it have to do with me?
Do you have the same questions as me before reading this book? What does math have to do with me? Mathematics seems to have nothing to do with us ordinary people except calculating the price of shopping and calculating the interest on saving money and loans. Or maybe your math still works elsewhere. So what does math have to do with me?
Three, this book is divided into several parts?
This book has *** 19 chapters, which are mainly divided into three parts. The first eight chapters are thinking articles, mainly about the influence of mathematics on thinking, reminding us to look at the world with rational thinking; The middle eight sections are methods, which mainly introduce the strategies and methods to solve difficult problems; The last three chapters are learning articles, mainly about how to learn and express suggestions.
In the part of thinking, the author tells us that we can look at the world with mathematical thinking from a mathematical perspective. In this part, I am deeply impressed by five points and want to share them with you. The first is the probabilistic worldview. We can't guarantee the final result of many things, but the probability of this result can be changed through our efforts. So what can we do? Accept the reality frankly and try to change the probability. Second, "prediction" is more important than "explanation". Why give an example of ceres' prediction? It's not that it's valuable and easy to "explain", but that it's valuable to predict, but it's really difficult. The third is the diversified dividend. "A team composed entirely of friends is very happy when discussing problems. These people have similar perspectives and opinions, so they are consistent most of the time. However, many of the conclusions they finally reached by synthesizing the opinions of all parties are wrong. Joining a stranger's team is different. Strangers and other people in the group live in different environments and have different perceptual perspectives. So the group discussion is full of arguments and differences. However, the final conclusion of the team is often correct. This is the dividend brought by diversification. " Just like an equation in mathematics, one angle is a result, and a set of equations can be obtained from multiple angles. Combined with the observation results from multiple angles, we can find the inner essence. Fourth, small happiness and great happiness. Which one would you choose? The author gives an example of buying a small house in the city center and a big house in the suburbs, and the commuting time is long. Finally, through a long list of mathematical formulas that I can't understand, he told us that frequent small fortune can bring more happiness than occasional great happiness. The fifth is about "advantages" and "disadvantages". It is also a series of mathematical formulas, which should be easy to read if you happen to understand them. But for me, a layman, although I can't understand the formula, I can at least understand that it can be deduced from a series of formulas by the author. "There are no advantages and disadvantages, only characteristics", which is scientific and scientifically proved.
Because there are too many formulas, most methods have been swallowed up, and only two things impress me the most. First, designers can learn more about the usage scenarios through verification and learning, and continue to develop this product according to the feedback from users. This will be very useful if you start a company to do product design. That is, "developing high-tech products cannot be done in one step." At first, we should make an imperfect but usable product, and then take it to the scene for use. After using it, engineers will tell us questions that we didn't expect at the beginning of the design, and users will ask us more. We will gradually improve on these bases. "The second is simulated annealing algorithm." Try randomness when you are young, but randomness should gradually decrease with age. When you are old and know what suits you best, you should control randomness, plough deep where you are best at, and don't change tracks easily. "
Although the study article has only three chapters, it makes me feel particularly useful. First, how to read books and newspapers to make the fastest progress? The emphasis is on forecasting and adjustment. "He will take the initiative to predict. When he saw a problem, he was not eager to see how others could solve it, but put forward a plan himself first. Compare your plan with the plan in the paper, learn from the gap and improve yourself. " "But no matter what kind of readers, good readers should choose to supervise learning, take the initiative to predict, and learn between readings. A good reader can adjust the speed at any time according to whether the prediction is correct or not. If the prediction is correct, it will be quickly swept away and slowly realize the mistake. This is active learning. " Second, universities study basic abilities and "learning methodology". The author inspires us from the intelligent machine learning mode that "the purpose of multi-task learning is to improve the average performance of the model on all tasks". As well as the transfer of learning, "students should pay attention to how to transfer the textbook knowledge they are studying at school to their future work." We should try our best to fully exercise these basic abilities, including comprehension, by studying every course. The third point is the incremental expression from primary to secondary. "I'll write a conclusion. Say important information first, and then gradually add details. "
Fourth, math has something to do with me.
The above thoughts on life, happiness and learning methods. All of them are demonstrated by corresponding mathematical formulas from a mathematical point of view. After reading this book, I suddenly found that mathematics has something to do with us ordinary people. This book helps us to re-examine our life and study from a scientific perspective. Although I haven't studied mathematics for many years, I have been skipping formulas I can't understand in the process of reading this book, but this book has opened a new perspective for me, a layman in mathematics, to observe the world, and I have benefited a lot from reading it. If you find it interesting, you might as well read it. I hope we can exchange different feelings.